HYPOTHESISE, VALIDATE, REPEAT - A CASE STUDY ON THE SITE OPTIMISATION FOR PORSCHE.COM/GB
In this case study, I am sharing how I conducted an initial exploratory research, proposed and led an optimisation strategy workshop for Porsche.com/gb, using quantitative data (Google Analytics), remote usability studies(UserZoom) and hypothesis building and validation.
Context
With the younger customers becoming more urban, they have become less interested in car ownership. Their primary touchpoint has been the web, and they were not used to visit car dealerships as much as their parents’ generation have been. There was an an increasing sense that the web (as well as the social media) would play a more significant role for the new emerging customer segments to learn about Porsche. However, Porsche GB did not know how the site performed and did not have a clear web strategy.
What needs to be optimised?
As a car manufacturer, the sales are mainly generated on on dealership site. It has long been unclear what Porsche.com/gb site actually aims to achieve.
So far, the key functions of the website have been widely understood as:
1) provide information on new vehicles
2) offer the car configurator platform
3) help customers to locate the nearest Porsche dealer
The current customer journey leads a customer who’s interested in purchasing a vehicle to explore the car configurator, and leave contact details so that they could be contacted by the dealership. This can be indeed an important KPI to track.
Is the current website fulfilling its potential to generate sales lead for Porsche Centres(Dealers) around the UK?
While asking around within the organisation, surprisingly, I have found out how the site actually performed to generate the leads has never been tracked studied within Porsche. My next step was to gather specific data to answer this question. I had to ask around to receive Google Analytics data, and received support from the Porsche Headquarters.
Today :
The conversion is low and the configurator has a room for further optimisation
Source: Google Analytics, July to August 2019
Google Analytics data from Porsche AG showed
1) A very low conversion rate
- only 0.4% of the total traffic convert leave contact details for the dealer at the end of the journey. Even considering the high value of the vehicle, this seems to be too low, if we look at 27% of the total visitors actually complete the configuration journey, which signals a serious interest in the product. At this stage, I’ve looked around what would be the industry benchmark for other car manufacturer’s sites would be, but it seems the automotive industry doesn’t seem to be very transparent to share such information. Google would know, for sure.
2) The car configuration journey is potentially extended
A big chunk of the car configurator users stay until ‘Options’ stage, but for ‘Accessory’ and ‘Summary, only 27% stay on, showing a steep decline.
Within the workshop, we’ve hypothesized 2 UX solutions would potentially solve these issues.
Move the log-in forward and make the users log-in before using the configurator
Simplify the car configurator, make stages easier to skip and provide options at which stages to leave the configuration process and let the Centres contact the user.
Building hypotheses
With this potential solutions in mind, we have built some assumptions on what results the new UX improvement must aim to achieve.
We believe that, if we
1) move the log-in process forward
2) make the car configurator journey shorter and easier,there will be more customers converting by the end of the journey.
In this scenario, by making log-in necessary for the customer to use the configurator, we hoped we would clear up cold leads early-on and keep hot leads to stay on. We’ve estimated at least 60% of the traffic to stay on to use the configurator, based on the current usage data on the configurator.
Regarding the configurator journey, by making the journey from ‘Interior’ on easier to skip (most users as we anecdotally know are primarily interested in the paint colours anyway), we made an estimate at the stage of ‘Dealer select’, at least 50% of users would stay on.
To optimise the site even further, we’ve made plans to experiment a small e-commerce feature - product suggestions on ‘Driver Selection’, which is a brand name for other Porsche merchandises, such as key chains and baseball caps. A small e-commerce purchase would be an opportunity for the customers to leave details for the test-drive, while collecting those goods within centres.
An hypothetical site performance after the UX changes
What research can validate this?
To learn how valid our hypotheses would be, I’ve recommended two different researches.
Designing the research to validate the hypotheses
At this stage, I’ve designed the research to understand
How the mandatory log-in would impact the final conversion of lead generation. This could be done with an A/B testing - launching two versions of site simultaneously on Optimizely.
If the new, simplified car configurator journey is as effective as/ more effective than the previous, longer journey in retaining the leads at the end of the configurator journey.
Once launched, the site with the new configurator would show the funnel performance on Google Analytics.